ELECTRICAL MONITORING OF PRESSURIZED INJECTIONS INTO A HEAP. Abstract

Transcription

1 ELECTRICAL MONITORING OF PRESSURIZED INJECTIONS INTO A HEAP Michael McNeill, Hydrogeophysics, Inc., Richland, WA Dale Rucker, Hydrogeophysics, Inc., Tucson, AZ Thom Seal, Metal Recovery Solutions, Inc., Spring Creek, NV Jeffrey Winterton, AngloGold Ashanti, Victor, CO Chris Baldyga, Hydrogeophysics, Inc., Tucson, AZ Jim Fink, Hydrogeophysics, Inc., Tucson, AZ Abstract Gold and copper mining of low grade ore typically involves an open pit circuit with a heap leaching facility. A heap consists of blasted rock placed in a large pile, which is subject to surface irrigation with a reagent to dissolve and transport the metal out of the rock. Pressurized injections (Hydro-Jex) into the heap leach pad at the Cripple Creek &Victor (CC&V) Gold Mine were also conducted to further enhance the recovery of gold. Hydro-Jex is operated by injecting fluid at very high pressures from a centralized well to targeted zones within the heap. Real-time electrical monitoring, where electrical current and voltage potential measurements were made using a 96 channel system to track the injections, provided both direct and indirect information to understand fluid movement during the injections. Although additional data processing would have provided valuable information, the strategy of focusing on the monitoring of time series data allowed for a quick turn-around. During this trial application of the Hydro-Jex technology, the timing of wetting front arrivals were noted in real-time by monitoring the electrical current output of the borehole electrodes; a sharp increase in current was a result of an increase in saturation, and hence a decrease in contact resistance with the heap. Flow direction was an important aspect during this project because the injections were near the slope of the heap and monitoring for stability of the rock pile was a necessary condition of the operating permit. Directionality of the wetting front was determined by mapping the shape of the apparent resistivity contours using a time-lapse mise-a-la-masse methodology. Overall, the monitoring showed no impact to the side slopes during injection and that the recovery of gold could be improved through simple redesign of the well field and injection timing. Introduction Heap leach operations are typically used in the retrieval of low grade gold, silver, and copper minerals. A heap is a rubblized rock pile constructed from blasted material, which is leached from the surface using common irrigation practices. A successful heap leaching operation can be dependent on many variables, including various interdependent phenomena such as ore placement, variations in ore type and grade, as well as the leaching efficiency (Rucker et al., 2009). For example, problems arising from excessive compaction or inadequate ripping of finer grained ore can lead to the development of preferential flow pathways, leaving areas of what are essentially trapped ore inventory in unleached portions of the heap. Minimizing the number and extent of these trapped ore inventory areas is of prime importance. A secondary recovery method for addressing such problems was recently developed (Seal 2004, 2007). The Hydro-Jex technology stimulates metal production by pushing solution deep within a heap from a centralized injection well, and directly to the underleached areas. Similar to enhanced recovery methods used in the oil and gas industries, solution is pumped at very high pressures in order to

2 overcome lithostatic pressure and open up new pathways for fluid to move into the heap. Typical Hydro-Jex operations can inject over 400,000 L of solution in less than two hours at depths that exceed 130 m below the surface. Understanding the direction of flow and area of influence of the pressured injections is important for optimizing recovery and ensuring safety of the side slopes. Few techniques exist that would allow large scale and rapid monitoring of such injections, with the exception of geophysically-based methods. In particular, those methods that are sensitive to the changes in saturation, such as electrical resistivity or ground penetrating radar, would allow monitoring of saturation changes. For this case, electrical resistivity is more easily applied for a large number of sampling locations. Geophysically-based electrical resistivity monitoring was used during a trial program of the Hydro-Jex technology at the CC&V Gold Mine in central Colorado (Seal et al., 2011). The monitoring was meant to complement the first three of nine injection sites by providing insight into the effectiveness of fluid movement into the heap. Due to the location of injection sites, which were quite close to the side slopes of the heap, environmental regulations also required that some type of monitoring be used to ensure slope stability during the injections. To satisfy these objectives, a combination of surface and borehole stainless steel electrodes, as well as the injection well itself, was used to make electrical measurements for real-time monitoring of the fluid movement. The usual geophysical processing and modeling methodology of inversion was not applicable in this case due to time restraints. Injections occurred rapidly, and information on slope stability and fluid coverage was needed in near real-time. A proper inverse analysis takes time to filter data, create input files, and run on slower field computers, even for truly optimized surveys. Therefore, electrical current output from the borehole electrodes was evaluated in real-time to track the wetting front of the injection flow; when the fluid encountered an electrode the current output would increase dramatically. These data, along with the voltage potential measurements, are presented to show the timing of wetting front arrivals through both the raw data and the analysis of processed apparent resistivity contours. The final results were then used to further optimize injections for future deployments of Hydro-Jex. Setting AngloGold Ashanti s CC&V Gold Mine is located in the historic mining district southeast of Cripple Creek, CO (Figure 1). Operations at CC&V consist of open-pit mining followed by heap leaching on a valley-fill-type heap that provides the sole means of gold production. The CC&V Gold Mine, in operation since 1994, has over time built up a very large Valley Leach Facility (VLF), now covering a lined area of 66.5x10 6 m 2. At the deepest point, the VLF is over 180 m deep. The growth of the VLF to over 200x10 6 megatons (mt) of ore is a testament to the historic productivity of the operation, but the extreme depths of the heap also create operational challenges that impact gold production from the heap. Drilling and sampling programs were introduced to directly inspect leaching conditions within the heap and assess reagent consumption and residual gold deportment. In 2008 and 2009, drilling samples retrieved from an area on the west side of the VLF showed both an abnormally high residual ore grade and a relatively low ph at over 100 m below surface. A review of ore stacking records indicated that approximately 1.5x10 6 mt of ore obtained from a historic dump had been placed in the affected area in The ore was essentially waste from a historic mine with grades sufficient to justify placing it on the leach pad, but it also contained a significant quantity of active sulfides that had been oxidizing for several decades. Very little lime was added to this ore when it was placed in 2001 and it was hypothesized that low alkalinity was severely limiting the efficacy of leaching in the area. Because

3 traditional surface-based secondary recovery techniques would be virtually ineffective in such a situation, Hydro-Jex was chosen for a trial program to investigate the viability of transporting additional alkalinity and the necessary reagents to recover this underleached ore. Figure 1: Location of Cripple Creek & Victor Gold Mine, west of Colorado Springs, CO. Hydro-Jex Seal (2007) describes the Hydro-Jex process as a three step procedure. In the first step, steel casings are driven down holes drilled in the heap and perforated at target depths corresponding to locations identified as having high concentrations of residual gold and/or adverse chemistry. Through a welded wellhead attached to the casing, a high-pressure pump is used to force solution down into the interior of the heap creating new solution paths while adding reagents to targeted zones of underleached ore. Figure 2 shows a photo of the injection skid with pump and connections to the well. A downhole packer system isolates each zone and controls the depth at which solution enters the heap, thus allowing specific areas to be targeted. Multiple depths can be targeted and treated by repositioning the packer system. Figure 2 also shows a schematic of idealized fluid pathways during injection. During the stimulation phase, any combination of reagents can be added to the solution pumped down the well to accelerate gold dissolution or remediate any adverse chemical conditions. The final stage of the process consists of periodically irrigating each well with leaching solution to rinse the dissolved gold to the liner and to further enhance chemical changes.

4 Figure 2: Left Illustration of the Hydro-Jex method showing four discrete injection zones. Right Hydro-Jex unit deployed at CC&V. Nine Hydro-Jex wells were drilled in the VLF ranging in depth from 116 m to 146 m. Figure 3 shows the site and the layout of the wells, with the initial estimated design diameter of fluid penetration approximately 50 m (represented by circles around each well location in Figure 3A). Each well had approximately 20 injection zones along its length. Three of the injection wells were monitored with electrical resistivity geophysics. All wells were monitored for pressure and flow rate. The actual penetration within each zone depended on site specific parameters such as depth, hydraulic resistivity, soil cohesion, and injection parameters. Figure 4 shows results of the injection procedure for two zones in HJ-9, with Figure 4A recording the injection data at a depth of 112 m and Figure 4B at 100 m depth. The data were recorded continuously with a datalogger and manually at the top of the well and included tophole pressure at the pump outlet, flow rate, and pump speed. The bottomhole pressure can be estimated by adding 11.2 bar to the pressures in Figure 4A and 9.8 bar to the pressures in Figure 4B to account for the depth and fluid friction loss. In Figure 4A, the initial flow before the pump is turned on was about 870 L/min at 5.5 bar line pressure (measured at the top of the borehole). At a few points along the timeline, the pumping is slowed in order to add lime slurry down the hole; pumping was resumed when the lime slurry had been offloaded. At the end of the injection cycle and after the pump is turned off, the flow is increased above the pre-injection values to around 3,020 L /min at 4.2 bar. If no mechanical changes had occurred in the zone such that the hydraulic conductivity (or hydraulic resistivity) was unaffected, then the pressure should have increased proportional to the increase in flow. However, given that the pressure dropped with an increase in flow, this indicates that the hydraulic conductivity increased due to the injections. In Figure 4B, cavity generation and fracture initiation is observed early in the injection cycle, before the first load of lime slurry was delivered. A drop in injection pressure with increased flow is a characteristic sign of fracture initiation at the critical pressure as described by Wu (2006). This pressure drop is observed multiple times indicating that the fracture actually causes new cavities to be formed before rupturing and fracturing again. During the extended injection period following the lime slurry additions, the flow of barren solution increases to around 4,150 L/min for about an hour while the pressure drops from about 9.3 bar to 8.9 bar.

5 Figure 3: (A, B, & C). Survey design for geophysical monitoring. A) Plan view map showing the nine Hydro-Jex injection well locations, B) Plan view map showing electrode layout for HJ-2 monitoring and C) 3-dimensional example of electrode layout. Electrical Monitoring Monitoring concepts A successful electrical monitoring system is dependent on the ability to transmit and receive electrical signal. A direct current is conducted into the ground through an electrode in contact with the earth, and subsequently a voltage potential is measured at a different location. Current output is proportional to the grounding, or the contact resistance, of the electrode. The level of contact resistance is a function of moisture; therefore an increase in moisture should result in a higher current output. Few have used the contact resistance to monitor the changes in subsurface conditions. Morris et al. (1996) and Lile et al. (1997) presented analyses of groundwater velocities from a tracer test using the method. Another technique for monitoring is the mise-a-la-masse method, which uses measured voltage originating from an electrical transmitter in contact with a fluid (or conductive body). In concept, the voltage contours can be used to track the fluid movement by mapping the equipotential lines at the

6 surface. Apparent resistivity values can be derived from these measurements and plotted in a colorcontoured map to estimate the shape of the conductive body (the injection wetting front). Figure 4: (A & B). Injection parameters for well HJ-9, with pressure and flow recorded continuously through a datalogger. Pump operation and lime slurry loads were recorded manually. A) Injection at 112 m and B) injection at 100 m below pad surface. MoL = Milk of Lime (i.e., lime slurry). Acquisition system The electrical resistivity data acquisition system used at CC&V was a trailer-mounted, multichanneled system capable of full 96 channel acquisition. In a 96 channel acquisition system, one electrode pair (the transmitter pair, or TX) passes electrical current into the subsurface to create an electrical field. The remaining electrodes are then used to measure the voltage potential on the other pairs (receiver pairs, or RX) simultaneously. The acquisition system then cycles through each pair for transmission. A data set consists of all electrode pairs taking turns at being the transmitter pair. For the Hydro-Jex monitoring, a pole-pole array was used, where one electrode from each of the TX and RX electrode pairs were placed effectively at infinity. The data acquisition system is composed of data acquisition cards, relay boards, transmitter boards, power supply, remote communication systems, and a PC used as the control computer. Software is used to control the TX and RX electrodes and is responsible for recording all data (including current

7 output and voltage input) in a database. A continuous series of data are stored for retrieval and processing. Once the injection has started, voltage and current data appear on the screen for real-time tracking of the injected plume. Survey design The design of the electrical resistivity monitoring was such that a series of electrodes were placed on the surface and within boreholes surrounding the Hydro-Jex well to capture the three dimensional propagation of the injected plumes. Specifically, the survey design for each of the three Hydro-Jex monitoring locations consisted of three borehole locations (with 14 electrodes per hole) and 48 surface electrodes. The surface electrodes were constructed of 18 long, 5/8 diameter stainless steel rods that were placed in direct contact with the heap material. Copper wire was used to connect the electrode to the acquisition system, and each electrode had its own dedicated wire. Borehole electrodes were constructed of wire bundles with individual wires running to an electrode. Fourteen electrodes were placed in each borehole, with a total of seven boreholes for the three injection wells. Electrode separation along the bundle was 10 m and the total length was 130 m. Figures 3B and 3C shows the electrode array design for the HJ-2 injection well. For any one set up around a Hydro-Jex well, three boreholes were used for a total of 42 electrodes. Installation occurred by dangling the electrodes in the cased hole, after which the casing segments were pulled to leave an open hole with the electrodes hanging at the surface. The boreholes were backfilled with heap material, with the goal of using moist, small, granular particles to help prevent bridging and ensure good grounding with the earth. An electrode that doesn t make direct contact cannot be used for the acquisition. Results Monitoring the movement of solution and defining flow characteristics during the injections were the primary goals for the project at CC&V. Arrays of surface and subsurface electrodes, designed to capture the three-dimensional propagation of the injected plumes through time, were installed around three of the injection wells to monitor the solution flow. In addition, given the nature of the project and the real-time data presentation to assure slope stability, the usual geophysical processing and modeling methodology of inversion, demonstrated in Rucker et al. (2009) and Seal et al. (2008), was not possible. Inversion modeling is a time-intensive process that aims to reconstruct the spatial distribution of the property electrical resistivity from all of the voltage measurement combinations. Instead, the electrical current output from the borehole electrodes was monitored for changes, which could be acquired without further processing. It was clear that an increase in saturation would reduce the contact resistance between the electrode and heap. This increase in saturation came from the injected fluid at the Hydro-Jex well. The main types of data recorded with the resistivity data acquisition system included current output on each electrode pair, voltage measurement on each electrode pair, and the range of errors expected for both. These time series data were used to directly monitor the injections. Figure 5 shows an example data set of electrical current, voltage (normalized to current and presented as the transfer resistance in ohms), and depth of injection. The electrical current data show definitive response to fluid injection, as values increase substantially during an injection (indicated in vertical bars). The voltage response on that same electrode (obtained during current transmission on the steel-cased Hydro-Jex injection well) also shows substantial change with injection timing and depth. When considered

8 spatially with all other pairs, the electrical data can reveal much about the position and direction of flow of the wetting front. Figure 5: Example data set during injection on HJ-2, showing voltage measurements and current transmission from a borehole electrode at a depth of 121 m. Figure 6 shows an example of the electrical current output from electrodes in borehole M3 during fluid injection on well HJ-2 (see Figure 3 for borehole electrode arrangement). Time series current data from three electrodes are shown, corresponding to depths of 60, 70, and 80 m below the surface of the heap. The injection schedule is also overlain on the plot to show the timing of wetting front arrivals at the electrodes relative to the start of each injection. For example, during injection at the 70 m depth, the electrodes at 80 and 70 m registered an arrival (red dot), during which time the current output almost doubled. The bottom portion of Figure 6 shows highly conceptualized plumes from the injection that could be interpreted from these arrival data. In most cases vertical drainage becomes the primary mode of solution movement. However, in other injection wells, it was noticed that the pressure from the injection was sufficiently high to overcome the effective stress and gravitational forces, evidenced by arrivals at electrodes at higher elevation than the injection. Throughout the remaining injections for the day the electrical current remained at elevated levels, suggesting that the saturation levels also remained high. These types of responses are not seen simultaneously on all three boreholes, indicating that plume migration is varying in each direction. Each evening, following the day s injections, a gravity fed fluid rinse was applied at an adjacent injection well (in this case HJ-3). A response noted on 82 (yellow dot) in the early hours of May 22, 2010 indicates that fluid from the HJ-3 overnight rinse reached this electrode. While the time series of electrical current provided a temporal assessment of wetting front arrivals, assessing the spatial distribution of the arrivals was accomplished by evaluating the voltage potential measured on the surface electrodes during current transmission on the centralized injection well. The voltages were linearly transformed to apparent resistivity (Rucker et al., 2010), compared to a background measurement just before injection began (as percent difference), and plotted as spatiallycontinuous contours.

9 Figure 6: Time series of electrical current during injection on HJ-2, using borehole electrodes in M3 as reference. The timing of wetting front arrivals is shown as a red dot. The bottom series of profiles show highly conceptualized views of the plume at the time of arrival based on the time series data. To demonstrate the outcome of an injection using time-lapsed mise-a-la-masse mapping, a forward resistivity model was conducted with a centralized well and a target. The target was modeled perfectly centered at the well with a diameter of 35 m and height of 5 m. The target was simulated with a resistivity of 1 ohm-m in a background of 100 ohm-m, using the resistivity model RES3DMODx64. The well was simulated as a vertical set of very conductive cells with a value of ohm-m and a diameter of 0.1 m. Figure 7 shows the results of the modeling, with Figure 7A showing the voltage contours around the well. Figure 7C shows the voltage converted to apparent resistivity. Figure 7E shows the apparent resistivity percent difference calculated from a background of no target. These series of plots show the contours to be symmetric around the injection well, and mimic the shape of the original target. However, when a similar target is off-center from the injection well, as shown in the series of plots on the right side of Figure 7 (7B, 7D, and 7F), the contours are not symmetric and do not

10 mimic the shape of the target. Figure 7F of the percent difference for the off-center target, reveals a high gradient at the center of the target and low gradient in the opposite direction of the target. Additionally, the location of the 0% difference contour in Figure 7F goes through the center of the target, making it a quantitative discriminator for geometric properties. Figure 7: Forward modeling of a simulated plume, showing voltage contours, apparent resistivity contours, and percent difference contours, as compared to a homogeneous background. The simulated plume (transparent gray circle) was modeled both centered and off-centered relative to the injection well (indicated by a +). Figure 8 shows spatial contour plots of percent difference in apparent resistivity for HJ-2 injection at 70 m. The percent difference was calculated from a baseline condition taken prior to any injections for the day (Figure 8A). The data in Figure 8 represent conditions during the second injection

11 for the day and voltage data were extracted in the middle of the injection (at 11:19 am) and near the end of the injection (at 12:57 pm) on May 20, The percent difference data are shown to range from a change of -4% to +0.3%, relative to the baseline. Figure 8B of mid-injection conditions shows a decrease in resistivity near the well and a slight increase away from the well. The gradient in the contours are the highest about 40 m to the west of the injection well, which could represent the center of the plume. Using electrical current data of Figure 6, wetting front arrivals at the M3 borehole occurred relatively early in the injection allowing us to fit the back side of the plume through those boreholes. The result is an elliptical wetting front that tends toward the slope of the heap. Hydrologically and geomechanically the propagation towards the slope is expected, given the lower pore pressures from reduced overburdened stresses. The heap is also less likely to be compacted, which could have significant impact on the permeability of the rock. Figure 8: Apparent resistivity data collected during HJ-2 injections at 70 m. A) Baseline apparent resistivity used for calculating percent difference plots. B) Percent difference contours for mid-injection at 70 m for HJ-2. C) Percent difference contours for post-injection at 70 m for HJ-2.

12 By the end of the injection (Figure 8C), the plume appears to have some lateral spread as interpreted from the decreasing percent difference values near the injection well and borehole electrodes. These interpretations, however, are purely speculative and do not provide the type of positional control compared to the electrical current data. Monitoring the contact resistance provides definitive benchmarking of the plume s wetting front, which can then be used in helping to interpret the apparent resistivity data. Regardless, the apparent resistivity data do give insight into the broader flow mechanisms than could not be obtained from contact resistance alone. Additionally, given the relative ease by which these data are processed and plotted, they allow real-time assessment of flow conditions in the heap. Conclusions A proof-of-concept injection experiment was conducted on a heap at the CC&V mine to help amend ph conditions and increase recovery deep within the rock pile. The injections were part of a technology called Hydro-Jex, which promotes mechanical changes in the heap to push solution far from the injection well. To more fully understand how far the solution may propagate, an electrically-based monitoring program was conducted in tandem to the injections, where series of electrodes were placed on the surface and within boreholes of the heap. A 96 channel acquisition system was used to rapidly acquire data during the injections, which typically occurred over 2-hour intervals. Given the rapid nature of the injection process and the need for real-time understanding of subsurface conditions, typical data processing using electrical resistivity tomography through inverse modeling was not conducted. Instead, data that could be discriminated with minimal processing were evaluated, such as the electrical current output on borehole electrodes and voltage data on surface electrodes during current transmission on the injection well. The electrical current output was proportional to the contact resistance each electrode had with the earth, which is a function of the degree of saturation. As saturation increased near the electrode, the contact resistance decreased allowing for a direct means for evaluating the arrival of the wetting front from the injection. These arrivals were then used to benchmark the location of the injection wherever borehole electrodes were located. A secondary method of processing included plotting the percent difference of apparent resistivity during the injection relative to background conditions. The apparent resistivity was calculated from voltage acquired on surface electrodes while passing current on the Hydro-Jex injection well. Background for differencing was established by extracting the data prior to any injection for a particular day. Forward modeling was also conducted to help interpret the resulting plots, which showed that offcenter plumes relative to the injection well will not have contours that mimic the shape of the plume. Instead, the center of the plume will be near the highest gradient of the contours and near the transition from negative to positive percent difference values. The results from the injections showed that this high gradient location was typically located west of the injection, under the side slope, where overburden stresses were minimized. The positions of an interpreted (and highly idealized) wetting front were verified through the electrical current data. Together, these datasets provided a real-time means of discriminating subsurface conditions without much ancillary processing. References Lile, O.B., M. Morris, J.S. Ronning, Estimating groundwater flow velocity from changes in contact resistance during a saltwater tracer experiment. Journal of Applied Geophysics 38:

13 Morris, M., J.S., Ronning, O.B. Lile, Geoelectric monitoring of a tracer injection experiment: Modeling and interpretation. Eur. J. Environ. Eng. Geophys. 1: Rucker, D.F., A. Schindler, M.T. Levitt, D.R., Glaser, Three-dimensional electrical resistivity imaging of a gold heap. Hydrometallurgy 98(3-4): Rucker, D.F., M. McNeill, A. Schindler, and G.E. Noonan, Monitoring of a secondary recovery application of leachate injection into a heap. Hydrometallurgy 99(3-4): Rucker, D.F., M.H. Loke, G.E. Noonan, and M.T. Levitt, Electrical resistivity characterization of an industrial site using long electrodes. Geophysics 75(4):WA95-WA104. Seal, T., Enhanced Gold Extraction in Cyanide Heap Leaching Using Hydro-Jex Technology, Ph.D. Dissertation, U of Idaho, May 2004 Seal, T., Hydro-Jex : Heap Leach Pad Stimulation Technology; Ready for World Wide Industrial Adoption? SME Annual Meeting and Exhibit Seal, T., Fink, J., Integrating Hydro-Fracturing Technology and Geophysics into 3-D mapping and Extraction of Metals in Heap Leaching; Hydro-Jex and High Resolution Resistivity. SME Annual Meeting and Exhibit 2008, paper number Seal, T., J. Winterton, D.F. Rucker, Hydro-Jex operations at AngloGold Ashanti s Cripple Creek & Victor Gold Mine. SME Annual Meeting and Exhibit Wu, R., Some fundamental mechanics of hydraulic fracturing. PhD Dissertation, School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia.